| From: | Merlin Moncure <mmoncure(at)gmail(dot)com> |
|---|---|
| To: | knizhnik <knizhnik(at)garret(dot)ru> |
| Cc: | Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com>, Oleg Bartunov <obartunov(at)gmail(dot)com>, Abhijit Menon-Sen <ams(at)2ndquadrant(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: In-Memory Columnar Store |
| Date: | 2013-12-12 18:26:47 |
| Message-ID: | CAHyXU0zDetDik7X8zEbSgYK3VncvVL_su3qrgSshVrJOtTyN_w@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Thu, Dec 12, 2013 at 12:18 PM, knizhnik <knizhnik(at)garret(dot)ru> wrote:
> IMHO it is strange to see such small default values in postgresql
> configuration.
This (low default work mem) is because of three things:
1) Most queries do not really need a lot of work mem
2) Work mem stacks with each query using it -- so with your 1mb
setting vs 1000 connections, you get a gigabyte. So, some
conservatism is justified although this setting tended to be much more
dangerous in the old days when we measured memory in megabytes.
3) Postgres does not query available physical memory for default
settings due to portability issues. So we tend to tune to "common
denominator".
merlin
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2013-12-12 18:29:40 | Re: ANALYZE sampling is too good |
| Previous Message | knizhnik | 2013-12-12 18:18:24 | Re: In-Memory Columnar Store |